HIV Testing for All Less Effective Than High-Risk Targeting

BALTIMORE, June 12 -- The old system of targeting high-risk persons for HIV testing is more efficient than screening all patients routinely, according to researcher here.

A targeted program -- using the same resources -- would identify more people infected without knowing it and avert more new HIV infections, asserted David Holtgrave, Ph.D., of Johns Hopkins Bloomberg School of Public Health.

On the other hand, because it would newly identify more HIV-positive people, the program would sharply increase the money needed for HIV care, Dr. Holtgrave reported in the online journal PLoS Medicine.

The opt-out testing program, proposed by the CDC last year, would make HIV tests a routine part of medical care, unless a patient specifically refused the test.

As proposed by the CDC, tests would take place without regard to a patient's risk of acquiring the virus and without the previous requirement of pre-test counseling.

The idea drew approval from American Medical Association but raised the hackles of The American Civil Liberties Union and the National Association of People with AIDS, both concerned about the lack of explicit consent.

But a key issue, Dr. Holtgrave said, is that no one has crunched any numbers to see how well such a program might perform.

The CDC estimates there are a million people with HIV in the U.S., about a quarter of them unaware they are infected.

The lack of knowledge is important, Dr. Holtgrave said, because studies have shown that people infected unwittingly with HIV transmit the virus at a rate of 8.8%, which drops to 2.4% if they become aware of their status.

Using mathematical tools, Dr. Holtgrave performed a cost-benefit analysis of four different scenarios over a one-year period, including the CDC's proposal, two variations of it, and a program in which the same million would be used to target, test, and counsel people in high-risk groups, such as injection drug users.

The analysis found:

• The CDC program and its variants would reach 56,940 people who were unaware that they were HIV-positive, compared with 188,170 for the targeted program.
• The CDC program and variants would avert between 3,644 and 5,333 HIV transmissions and new infections (depending some assumptions about risk behavior), compared with 14,453.
• The cost per transmission or infection prevented would range from ,149 to ,993 for the CDC program and variants, compared with ,383 for the targeted program.
• The public support needed for HIV care for the newly diagnosed would be million for the CDC program and variants, but would reach ,176 million for the targeted program.

There are several limitations to the analysis, Dr. Holtgrave said, starting with the fact that many of the parameters he used -- such as the number of people with HIV who don't know it -- are not known with any degree of precision.

But the value of the study, he said, is that it allows people involved in HIV care and prevention to sharpen their estimates of how HIV testing should be carried out and it identified areas in which more research is needed.

"While the CDC's recommended opt-out testing offers some public health benefit, the data shows there would be substantially more benefit from a more targeted program that includes rather than discards risk reduction counseling -- including more diagnosed infections and more transmissions prevented," he said.

The study is "provocative," agreed Ronald Valdiserri, M.D., of the VA, who was formerly at the CDC and involved in developing the new testing guidelines.

But, writing in an accompanying perspective article, he said it "would be a mistake" to reduce the debate to an either-or discussion of the CDC's plan or targeted testing of high-risk populations.

The important thing is to catch HIV early, both from a patient and a public health viewpoint, and "perspectives on how best to encourage early diagnosis of HIV infection will continue to evolve," Dr. Valdiserri said.

In an Annals editor's summary, it was pointed out that this study has a major limitation in that it tried to predict what might happen in the future. "It did not study the actual impact of the two different types of testing on a group of people," he summary noted. "Studies such as this one, which try to predict t he future, are always based on a number of assumptions and these assumptions may turn out not to be true."